Nicolas Ragot

WRITER STYLE ADAPTATION IN ONLINE HANDWRITING RECOGNIZERS BY A FUZZY MECHANISM APPROACH: THE ADAPT METHOD

This study presents an automatic online adaptation mechanism to the handwriting style of a writer for the recognition of isolated handwritten characters. The classifier we use here is based on a Fuzzy Inference System (FIS) similar to those we have designed for handwriting recognition. In this FIS each premise rule is composed of a fuzzy prototype which represents intrinsic properties of a class. Furthermore, the conclusion part of rules associates a score to the prototype for each class. The adaptation mechanism affects both the conclusions of the rules and the fuzzy prototypes by recentering and reshaping them thanks to a new approach called ADAPT inspired by the Learning Vector Quantization. Thus the FIS is automatically fitted to the handwriting style of the writer that currently uses the system. Our adaptation mechanism is compared with well known adaptation techniques. The tests were based on eight different writers and the results illustrate the benefits of the method in terms of error rate reduction (86% in average). This allows such kind of simple classifiers to achieve up to 98.4% of recognition accuracy on the 26 Latin letters in a writer dependent context.

A generic hybrid classifier based on hierarchical fuzzy modeling: experiments on on-line handwritten character recognition

In our previous works, a recognition system named ResifCarwas designed specifically for on-line handwrittencharacter recognition. This system is based on an explicitmodeling by hierarchical fuzzy rules. Thus, it is understandablean optimizable after the learning stage. We present inthis article a new classifier that is an extension of ResifCar.Indeed it tries to combine ResifCars advantages witha generic aspect to handle different recognition problems.This new hybrid system combines two complementary levels.The first one uses a robust modeling by an intrinsicfuzzy clustering of each class and determines their confusingareas. The second level, based on fuzzy decision trees,operates a progressive discrimination inside these areas.Both levels are formalized by fuzzy inference systems organizedhierarchically and fused for final decision. Experimentswere conducted on the one hand on classical benchmarksand on the other hand on on-line handwritten digitsand lower-case letters. For all of these cases, the classifierachieves good recognition rates without final optimization.

A new hybrid learning method for fuzzy decision trees

This paper presents a new hybrid learning method for the construction of fuzzy decision trees. The main principle of this approach is to automatically generates a hierarchical organization of the knowledge coupled with local choice of the best feature subspace. To improve the representation, a double level of modeling is used. Firstly a pre-classification level searches fuzzy decision regions to operate a natural discrimination between classes. The second level refines the previous one, doing an intrinsic fuzzy modeling of the classes represented in the fuzzy regions. Moreover, the best feature subspace is determined locally by a genetic algorithm for each partitioning. Finally, to have an understandable and transparent representation, the fuzzy decision tree is formalized as a fuzzy inference system which is easily modifiable and can be optimized a posteriori. First experimental results conducted on classical benchmarks and on a handwritten digits database show the capacity of the hybrid learning approach to provide reliable and compact classification system.

Système de classification hybride interprétable par construction automatique de systèmes d'inférence floue

En reconnaissance des formes, il est nsouvent difficile de concevoir des systemes a la fois performants, generiques net interpretables. Pourtant, linterpretabilite du nsysteme permet a un expert de le nmaintenir et doptimiser ses nperformances (taux de reconnaissance, nfiabilite, occupation memoire, ...). Afin de repondre a ces nobjectifs, nous proposons un nouveau classifieur hybride qui repose nsur une combinaison originale de deux niveaux nde modelisation complementaires bases sur des connaissances nintrinseques et discriminantes. Les informations sont modelisees de nfacon robuste et explicite par des sous-ensembles flous extraits de maniere nautomatique. Des systemes dinference floue permettent alors dagreger net de fusionner les connaissances afin dobtenir un nprocessus de decision robuste et comprehensible. Les nexperimentations reportees dans cet article valident aussi nles proprietes de genericite et de performances.

The Prototyping and Focused Discriminating Strategy for Pattern Recognition and one Instantiation: the MELIDIS System

This paper presents the Prototyping and Focused Discriminating (PFD) strategy for pattern recognition. This strategy takes benefits from the duality between model generation and discrimination. Both collaborate through a focusing mechanism that detects the conflicts between the class models and drive the discrimination. Classifiers based on this collaboration benefit from a set of useful properties. The Melidis system illustrates this strategy and extends its possibilities, using a fuzzy framework. As shown by experiments, the resulting system provides an interesting compromise between accuracy and compactness. Experiments also demonstrate the interest of the new strategy and of its focusing mechanism.